Key Technology Research Of Ontology Mapping For The Semantic Web

In recent years, the ontology as a semantic-based knowledge representation model iswidely used in various fields, such as knowledge representation, knowledge sharing andknowledge integration. At present, we do not have a unified standard for the construction ofontology. With the rising number of ontology, it will inevitably generate a lot of ontologywhich have different structure but belong to the same field. When communicate between theheterogeneous ontology, the inconsistent of representation and internal logic will produceblock on knowledge sharing and reuse. In order to accomplish the reuse and communicate ofknowledge between heterogeneous ontology, need to find a mapping relationship betweenthem, the process of mapping discovery is called ontology mapping.Ontology mapping is the key technology to solve the heterogeneous problem amongontology, has great significance on realize the communication and interoperation betweendifferent systems. As the ontology-object has a rapidly increasing on the size and quantity, iturgent needs to research some methods that can be (semi) automatic accomplish a largenumber of ontology mapping.This article in view of the existing large amount of heterogeneous ontology, design andimplementation a kind of quick and accurate method to realize ontology mapping. Firstly, thispaper fully studies the various features of WordNet, finds that the number of shared semanticpaths have a significant impact to the calculating of similarity, and proposes a new semanticsimilarity calculation method that based on WordNet shared path which is the most importantmethod on ontology mapping. Using classical dataset to test proved that it has obviousadvantages. Secondly, comprehensive utilization of characteristic information of ontology,design and implementation an improved method involves some feature information, such asthe name of ontology concept, context, structure, properties and instance. These methods canfurther improve the precision. Finally, propose a framework which can accurately build arelationship in semantic with two heterogeneous ontologies. In this framework, not only usethe semantic information of ontology concepts, but also consider the context information, thusto further improve the accuracy of the mapping discovery. Using OAEI2012experimentalcases to test its effectiveness and respectively through the recall, precision and F test value tomeasure the comprehensive performance of the framework. Experiments show that mapping based on semantic feature can greatly improve its accuracy.